Just an anecdote, to tuck away in case you ever need it.
 
I have a 256GB SSD on my internal SATA bus, which I use for my (Parallels) virtual machines. Today I decided to update Windows 10, and away we went, on a long download because I foolishly opted for developer builds a year or so ago. It finally downloaded and started the 20-minute update process, and got about 25% thru the install when suddenly the SSD dropped of the bus. (That means it unmounted and disappeared.) Not good. Needless to say, the VM image was toast, but I have backups, so no problem.
 
So I restored and tried again… and got pretty much the same thing, albeit later in the update process.
 
Fine… off to re-initialize the drive. It was initialize with SoftRAID, so I headed over there to reformat it… and it couldn’t get pas 1/3 way without tossing up I/O errors.
 
Not good, again. I tried another SSD… and got the same thing! I tried a different SATA port (still internal) and got the same thing.
 
Now I’m surmising that it wasn’t the SSD nor the port used. Off to SMARTUtility, to see if it would tell me anything.
 
Yep. Failing. (Really? an SSD with less than a year on it?)… but -also- out in an external port-multiplied enclosure with two other drives, was my TimeMachine drive, a Seagate 3TB I tossed in to see if I could live with TimeMachine yet. And it was marked (by SMARTUtility) as “failed”.
 
WHEE. Are we having fun yet?
 
Well, first: the Seagate 3T drive has a higher failure rate than pretty much all other drives, (almost “all other drives combined”) so I decided to yank it, and see what happened.
 
What happened was that as soon as that Seagate 3T was out of the mix, (and even though it was controlled by a PCIe card, and not the motherboard SATA) the SSD ran like a champ.
 
So, to be sure, I stuck the drive into the computer (the motherboard port) and it had massive I/O errors, and would not format at all. (I destroyed it, and put the parts in the trash.)
 
What makes this interesting is that one failing drive, on an entirely different bus, caused an SSD on another bus to appear as failing. Removing the truly malfunctioning drive, fixed the situation.
 
(Oh… and the Seagate 3T had less than 3000 hours on it. If the cause-and-effect of this little story doesn’t interest you, perhaps the recommendation to stay away from Seagate 3TB drives will … )